On Alpha processors, converts the expression on the right-hand side of the equal sign to the IEEE T_floating type (double precision, length 8 bytes) and deposits the result into the specified location.